mad-dev March 2002

mad-dev@lists.mars.org

19 participants
27 discussions

Recalculate CRC checksum
by Dirk Nehring 31 Mar '02

31 Mar '02

Hi, I'm writing a program to correct bad MP3 files (by dropping bad frames, dropping IDv2 tags on demand, etc.) by using libmad. One point is that I want to add an CRC-16 checksum to the header. Can I use the libmad low level interface for this (i.e. set protection bit, and call mad_frame_decode(), get crc-checksum from header->crc_target)? Any ideas? Dirk

2 3

configure scripts and cross-compiling
by Tripp Lilley 29 Mar '02

29 Mar '02

In many places, the configure scripts do tests like this: case "$host" in ... Should those be changed to: case "$target" in ... ? I find, in trying to cross-compile for arm-linux from i686-linux (for the Rio Receiver), that the configure doesn't "do the right thing", which is what leads to the fix-make script in Jeff Mock's patch here: http://www.mock.com/receiver/demo/ I'd rather get the configure scripts "corrected" (if I am, in fact, right about the host vs. target bit) in the sources :) While I'm at it :) what would it take to get Jeff's Rio Receiver audio device patches applied? Thanks!

2 1

Re: [mad-dev] MAD Licence and Static link
by armin.gerritsen＠philips.com 28 Mar '02

28 Mar '02

>You forgot the most important one: to qualify for this exemption, you >must "distribute [the non-GPL linking parts] as separate works." OK, but that is also no problem for me. The non-GPL part will be distributed only through and by the company and the generic part that uses MAD will be on some public non-company related server available to the whole world. Thanks for claring this up. I think I'm going to wrap the MAD in a GPL DirectShow COM component which will use an as generic possible setup (which for those who know DirectShow know, means basicly using standard MEDIATYPE definitions). That way there is no doubt about it that it is indeed independant and seperate from any other program ever using it. (So something similar to what they did with LAME.) Regards, Armin

1 0

MAD Licence and Static link
by armin.gerritsen＠philips.com 28 Mar '02

28 Mar '02

Hello, I want to use MAD in a commercial product (COM component) on Windows and Windows CE. I understood, that if I link MAD as a seperate library (DLL), I would not break GPL. However, now I was wondering if I would link MAD staticaly, but make sure all calls are in one seperate source-file and distribute with the product that seperate source-file together with all documentation related to that file, giving credit to MAD, etc., and keep the rest of the source closed, I would still not break GPL. I'm actually only using MAD in the body of a function as in the following pseudo code: STATUS DecodeBuffer( inbuffer, outbuffer ) ; { /* Using MAD to decode the buffer */ } Maybe I get my company to make the whole product GPL, but I'm afraid that they'll not agree. :-) Using a DLL is of course an option, but I'd prefer to distribute just one object. But I definitly don't want to break GPL or so. Regards, Armin PS When will MAD support MP3pro? The MP3pro files I decode now, are played at too 'high speed'. I seems to get the bitrate incorrect.

4 3

Re: mad-dev digest, Vol 1 #136 - 2 msgs
by armin.gerritsen＠philips.com 27 Mar '02

27 Mar '02

>If you use MAD as a DLL (not sure about the COM object), you have to distribute >your product under the GPL, too. This is not the Library (or lesser) GPL! The >same is for static linking. > >The complete license can be found under >http://www.gnu.org/copyleft/gpl.html Aha, I see there: "... If identifiable sections of that work are not derived from the Program, and can be reasonably considered independent and separate works in themselves, then this License, and its terms, do not apply to those sections when you distribute them as separate works. " So I need to make my sure the non-MAD part is - identifiable as sections of that are not derived from MAD - reasonably considered as an independent and separate work in itselve Part 1 is no problem. Part two is a problem when I link statically, even if I give that one source-file free. This means I do have to use my seperate source-file that uses MAD, as a independant module (DLL or even its own COM object as they did with LAME), which I need to make completely GPL too. I can link MAD statically in that DLL of course. That takes care of the 'seperate' part. And if the setup of that MAD-module object (DLL or COM or whatever) is generic enough to be considerred 'independant' from my COM object (in the sence that it is not *just* created for my own COM object), I don't have to give the COM source-code free. This shouldn't be that hard since I'm planning on re-using this MAD-based module in several other completely different projects too. So no static linking. ;-( >I always thought MP3pro is backward compatible? It is, but MAD identifies the bitrate incorrect in some cases. So it is a MAD problem. It is a comon problem that other decoders sometimes also have. Regards, Armin

1 0

MACC instruction
by Gabriel Bouvigne 25 Mar '02

25 Mar '02

Could someone explain why the multiply and accumulate instruction seems so important in Mad? Regards, ---- Gabriel Bouvigne www.mp3-tech.org personal page: http://gabriel.mp3-tech.org

2 1

Decoding partial files (1st 30s)
by iroro 22 Mar '02

22 Mar '02

Hi maddev-group & Rob, I'd like to know the easiest way to decode partial files, for example only the first 30 seconds. for constant bitrate layer 3 encodings, the first 30 seconds can be calculated by time = (size * 8) / (bitrate * 1000)? [from the dev-archive you posted in response to Joe Drew's post: Seeking, frames and time] but to get the bitrate, you have to read @ least one frame and from them can calculate the # of frames necessary for 30seconds ? to make things more tricky ... in reading the CBR layer 3 frames you have to read into the subsequent frames to really know if a current frame has ended or not. I know most of the incremental file descriptor reading action happens "decode_input_read()" ...so if i manage to calculate the # of Frames that correspond to 30 seconds of pcm, what structures do i need to set to stop decoding after the #of frames has been reached? MAD_FLOW_IGNORE? so the more general question is what is the best way for CBR and VBR layer 3 encodings to decode just the first 30 seconds to pcm? especially without having to scan through the whole file ... thanks a lot, iroro -- --o0o------------------------------------------------------o0o-- minimis optime peractis, maxima bene agentur --o0o------------------------------------------------------o0o--

2 1

Re: [mad-dev] how build a standalone mad for arm?
by acen10 19 Mar '02

19 Mar '02

hi Gareth, now i can run minimad and it output file,but what is the output file's format? it may be inputed directly to DAC? the madplay can output raw format file and it can play when i simply add wave format header to it. but minimad'output dont the same as madplay's. On 2002-03-18 11:38:00 Gareth wrote: >> i dont know how to initiate arm and set the start address >> using arm-linux.whether in ctr0.s ? > >I think the function 'decode' will just take a pointer to where the mp3 is. >In minimad.c : decode(fdm, stat.st_size); > >where fdm is the pointer to the mp3 data > >> what do the stdin and stdout refer to ? keyboard and screen? >> in minimad.c : >> int main(int argc, char *argv[]) >> { >> struct stat stat; >> void *fdm; >> >> if (argc != 1) >> return 1; >> .. >> if you run minimad with any parameters it will return doing >> nothing. > >Yes it will normally print the data to the screen so you redirect it to a >file using "> audio.pcm." > >"if (argc != 1)" means return if the arguments are not 1. So if there are no >arguments it will do nothing. > >Have you managed to compile minimad.c for arm yet? Best regards. acen10 acen10(a)263.net

2 1

Re: Re: [mad-dev] how build a standalone mad for arm?
by acen10 18 Mar '02

18 Mar '02

On 2002-03-16 12:22:00 Gareth Bransby wrote: >> the lartware's compile-tool is arm-linux ,the executable >>can >> run without linux? i think it should be arm-elf if i wanna build >> standalone mad,that is right? > >The arm-linux compiler should be fine as far as i know. As >long as your code has no system calls it should run without >linux. For example minimad.c has things like that mmap >call, which i think requires an os. > i dont know how to initiate arm and set the start address using arm-linux.whether in ctr0.s ? >> minimad can >> be compiled and ran no complaint,but it look like doing nothing. > >Use minimad like this : ./minimad < file.mp3 > file.wav >It reads and writes from stdin and stdout. > what do the stdin and stdout refer to ? keyboard and screen? in minimad.c : int main(int argc, char *argv[]) { struct stat stat; void *fdm; if (argc != 1) return 1; .. if you run minimad with any parameters it will return doing nothing. > acen10 acen10(a)263.net

2 1

Re: [mad-dev] how build a standalone mad for arm?
by acen10＠263.net 16 Mar '02

16 Mar '02

Hi, On 2002-03-15 10:29:00 Gareth Bransby wrote: >Hi > >I am attempting the same thing as you. I have compiled libmad using the cross compiler from the Lart website : > >www.lart.tudelft.nl/lartware/compile-tools/ the lartware's compile-tool is arm-linux ,the executable can run without linux? i think it should be arm-elf if i wanna build standalone mad,that is right? >I think minimad.c can be modified fairly easily so it is >independent of any os or system calls. I am having trouble >cross compiling it though so i can start changing it. My >board only has an ARM7td - no enhanced multiply. i am using cygwin and i configured and built successfully mad-0.13.0b on my host.the madplay can produce wave file which can be playback .minimad can be compiled and ran no complaint,but it look like doing nothing. would you like tell me any idea to modify minimad.c? acen10 acen10(a)263.net

2 1

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

mad-dev March 2002